The ability of a robotic hand to grasp an apple without crushing it heavily depends on its intrinsic mechanics and not just visual recognition by AI. The prior probability that AI correctly identifies the apple is high, estimated at about 90%, but this does not guarantee a delicate grip. For an apple to remain intact, the conditional probability that force sensors are well-tuned is crucial, say at 85%. If, for example, Festo's hand lacked adequate pressure sensors, even perfect AI recognition would likely cause damage, reducing success probability to less than 20%.